Vector
High-performance observability data pipeline. Sources, transforms, sinks — written in Rust by Datadog (open source).
| License | MPL-2.0 |
| Repository | https://github.com/vectordotdev/vector |
| GitHub stars (approx) | 22000 |
| Language | Rust |
| Self-host deployment | single binary, docker, kubernetes daemonset |
| Commercial hosted option | no / self-host only |
| Freshness status | Fresh |
| Freshness checked | 2026-05-17 |
| Categories | logs, metrics |
Vector is the logs/metrics ETL layer most teams discover when they outgrow Fluentd or Logstash. Rust, single binary, hundreds of source/sink integrations, VRL (Vector Remap Language) for transforms.
Pick it if: you need a fast, memory-efficient agent or aggregator to reshape, filter, sample, or fan-out telemetry before it hits storage. Excellent for cost control — drop noisy logs at the edge.
Skip it if: you only have Kubernetes container logs and Fluent Bit is already working (Fluent Bit is lighter for that single use case). Also skip if you need OpenTelemetry traces — Vector is logs+metrics, the OTel Collector is the better trace pipeline.
Deployment: binary on hosts, sidecar in k8s, or daemonset. Note: Datadog owns the project; the license is MPL-2.0 and development remains active and open, but the governance is single-vendor.